Heart failure patients often exhibit dyssynchronous ventricular contraction due to conduction abnormalities. Cardiac resynchronization therapy (CRT) aims to re-coordinate contraction by applying appropriately timed pacing stimuli to the failing ventricles. CRT reduces morbidity and mortality, yet approximately 30% of patients fail to respond to CRT. Current dyssynchrony indices used to identify potential responders to CRT, which are based only on the local electromechanical activity of the heart, have poor predictive capability, demonstrating an incomplete understanding of the electromechanical behavior of the failing heart. The overall objective of this research is to characterize the relation between the electrical and mechanical activity in normal and failing hearts under different loading conditions. This relates to the NHLBI mission to support basic research that investigates the causes and treatments of heart disease. To achieve the objective of the proposed research, I intend to develop and validate, from magnetic resonance imaging (MRI), diffusion tensor (DT) MRI, electrophysiological recordings and ventricular deformation data from MRI tagging, 3D anatomically-accurate computational models of ventricular electromechanics in normal and failing canine hearts. These models will be used to test the hypothesis that the 3D distribution of the electromechanical delay (EMD) is heterogeneous in both normal and failing hearts and is altered by the loading conditions. In addition, I surmise that the deleterious remodeling due to heart failure results in extended EMD, particularly at the late-depolarized regions, and increased heterogeneity in the 3D distribution of EMD. The specific aims of this study are as follows: 1) Develop and validate, from MRI, DTMRI, electrophysiological recordings and ventricular deformation data from MRI tagging, detailed high-resolution 3D anatomically-accurate electromechanical models of normal and failing canine hearts. 2) Using the electromechanical model of the normal canine ventricles, determine the 3D distribution of EMD under different loading conditions. 3) Using the electromechanical model of the failing canine ventricles, determine how the detrimental electromechanical remodeling resulting from heart failure alters the 3D EMD distribution in the canine ventricles for the same loading conditions as in Specific Aim #2. The development of a validated realistic model of ventricular electromechanics, as proposed in this application, overcomes the inability of current experimental techniques to simultaneously record the 3D electrical and mechanical activity of the heart with high spatiotemporal resolution. The new insights into the electromechanical behavior in the normal and failing ventricles under different loading conditions to be acquired under this study are expected to ultimately lead to rational optimization of CRT delivery and improvement in selection criteria for identifying viable CRT candidates.